j-w-yun/optimizer-visualization
An educational visualization tool that animates the convergence paths of TensorFlow optimizers on loss surfaces.
Not currently ranked — collecting fresh signals.
star history
This repository creates animated visualizations comparing how different TensorFlow optimizers traverse loss landscapes toward minima. Users can observe how algorithms like Adam, Adadelta, Adagrad, Momentum, RMSProp, and GradientDescent behave differently when gradients are steep or near-flat. It serves as a learning aid for understanding optimization algorithm dynamics.
Frequently asked
- What is j-w-yun/optimizer-visualization?
- An educational visualization tool that animates the convergence paths of TensorFlow optimizers on loss surfaces.
- Is optimizer-visualization open source?
- Yes — j-w-yun/optimizer-visualization is open source, released under the MIT license.
- What language is optimizer-visualization written in?
- j-w-yun/optimizer-visualization is primarily written in Python.
- How popular is optimizer-visualization?
- j-w-yun/optimizer-visualization has 402 stars on GitHub.
- Where can I find optimizer-visualization?
- j-w-yun/optimizer-visualization is on GitHub at https://github.com/j-w-yun/optimizer-visualization.